Lead Data Analyst · AdTech · Experimentation · Analytics Engineering
I own analytics problems end to end: framing the business question, designing the methodology, and shipping the result as a production tool instead of a one-off report. At one of the largest tech companies in Eastern Europe I have built six Python libraries, SDKs, and services that analysts and engineers from several teams now run in their own pipelines.
| Project | What it does | Package |
|---|---|---|
| fast_exp_analytics | A/B and A/B/C experiment analytics: significance tests, MDE, power analysis, duration planning, Excel reports, chat-ready summaries |  |
| anomaly_impact_alert | Anomaly detection, forecasting, deviation decomposition, Telegram alerts. Z-Score, STL, SESD, LOF, Isolation Forest, Prophet, ETS. Colab demo included |  |
| ytsaurus_python_client | Python client for YTsaurus, YQL, and CHYT: large results into pandas, async long-running queries, DataFrame uploads |  |
| ytsaurus_clickhouse_proxy | FastAPI proxy that exposes YTsaurus CHYT as a ClickHouse endpoint, so DBeaver and DataGrip connect over JDBC with a logical table catalog |  |
| streamlens | Live streams in one auto-updating Telegram message: viewers, trends and a slideshow of frames for YouTube, Twitch and Kick. Bring your own data or let it watch channels; no database, CLI and Docker included |  |
| KlipRun Desktop | macOS app for OpenCode and Codex sessions: embedded terminals, Kanban and Office views, usage statistics, and notifications |  |
Also: bot_streams_sender (the original Airflow + Postgres live-stream tracker with anomaly detection) · gym_online_checker (gym occupancy heatmaps to Telegram).
- Built an anomaly monitoring platform for advertising metrics: 7 statistical and ML detectors, expected-value forecasting, deviation decomposition by product and segment, automated alerts to responsible analysts. Scaled it to 264 production scenarios orchestrated as a generated 530-node workflow with retries, backfill, and ownership routing.
- Created an internal experimentation library for A/B and A/B/C tests: statistical tests for different metric types, confidence intervals, MDE, power analysis, duration planning, multiple-testing corrections, automated reporting. Moved regular experiment analysis from notebooks to production pipelines.
- Developed a data access SDK for YTsaurus / YQL / ClickHouse, used in 70+ production scripts and by analysts across several teams; recommended in Data Engineering channels. Set up Data Quality monitoring for analytical tables: freshness, missing partitions, volume deviations.
Estimated new ad product opportunities with scenario models (adoption, budget redistribution, cannibalization); the results defined the MVP segment and validation criteria.
- Mentor junior and middle analysts, run onboarding, and hold a weekly technical interview section.
- Run internal meetups and workshops: product analytics, mobile attribution, Git, Python packaging, CLI tools, AI agents.
- Brought AI coding agents and MCP integrations into the team's workflow: data exploration, code review, documentation, corporate systems.
Python · SQL (ClickHouse, YQL, Postgres) · YTsaurus · pandas, SciPy, statsmodels, scikit-learn, Prophet · Airflow · FastAPI · Docker · GitLab CI · pytest · Superset, Jupyter
📬 Open to conversations about analytics, experimentation, and data tooling: Telegram · LinkedIn